INDEX
Explanations
time references related to years
New Auto-Interp
Neuron Alignment
Index
Value
% of L₁
50
+0.25
1.3%
1053
+0.11
0.6%
11
+0.11
0.5%
Correlated Neurons
Index
P. Corr.
Cos Sim.
1363
+0.25
0.05
577
+0.11
0.04
776
+0.11
0.05
Negative Logits
<bos>
-2.51
ⓧ
-1.04
<?
-0.91
/**
-0.75
jadx
-0.73
-0.71
<?
-0.63
Datuak
-0.60
lateinit
-0.59
ɵɵ
-0.58
POSITIVE LOGITS
bandung
1.20
sergio
1.06
jorge
1.05
Palembang
1.04
jati
1.03
santiago
1.02
rodriguez
1.02
lorenzo
1.01
ricardo
1.01
jawa
1.00
Activations Density 0.095%